Multiagent Systems is the title of a collection of papers dedicated to surveying specific themes of Multiagent Systems (MAS) and Distributed Artificial Intelligence (DAI). All of them authored by leading researchers of this dynamic multidisciplinary field.

/pdf/multiagent-systems-a-modern-approach-to-distributed-5fon0donq1.pdf

Multiagent systems: a modern approach to distributed artificial intelligence

Supervised machine learning algorithms have been a dominant method in the data mining field. Disease prediction using health data has recently shown a potential application area for these methods. This study aims to identify the key trends among different types of supervised machine learning algorithms, and their performance and usage for disease risk prediction. In this study, extensive research efforts were made to identify those studies that applied more than one supervised machine learning algorithm on single disease prediction. Two databases (i.e., Scopus and PubMed) were searched for different types of search items. Thus, we selected 48 articles in total for the comparison among variants supervised machine learning algorithms for disease prediction. We found that the Support Vector Machine (SVM) algorithm is applied most frequently (in 29 studies) followed by the Naive Bayes algorithm (in 23 studies). However, the Random Forest (RF) algorithm showed superior accuracy comparatively. Of the 17 studies where it was applied, RF showed the highest accuracy in 9 of them, i.e., 53%. This was followed by SVM which topped in 41% of the studies it was considered. This study provides a wide overview of the relative performance of different variants of supervised machine learning algorithms for disease prediction. This important information of relative performance can be used to aid researchers in the selection of an appropriate supervised machine learning algorithm for their studies.

/pdf/comparing-different-supervised-machine-learning-algorithms-3azrndzb2b.pdf

Comparing different supervised machine learning algorithms for disease prediction

As computing becomes pervasive, embedded systems are deployed in a wide range of domains, including industrial systems, critical infrastructures, private and public spaces as well as portable and wearable applications. An integral part of the functionality of these systems is the storage, access and transmission of private, sensitive or even critical information. Therefore, the confidentiality and integrity of the resources and services of said devices constitutes a prominent issue that must be considered during their design. There is a variety of cryptographic mechanisms which can be used to safeguard the confidentiality and integrity of stored and transmitted information. In the context of embedded systems, however, the problem at hand is exacerbated by the resource-constrained nature of the devices, in conjunction with the persistent need for smaller size and lower production costs. This paper provides a comparative analysis of lightweight cryptographic algorithms applicable to such devices, presenting recent advances in the field for symmetric and asymmetric algorithms as well as hash functions. A classification and evaluation of the schemes is also provided, utilizing relevant metrics in order to assess their suitability for various types of embedded systems.

Lightweight Cryptography for Embedded Systems A Comparative Analysis

To collect software bugs found by users, development teams often set up bug trackers using systems such as Bugzilla. Developers would then fix some of the bugs and commit corresponding code changes into version control systems such as svn or git. Unfortunately, the links between bug reports and code changes are missing for many software projects as the bug tracking and version control systems are often maintained separately. Yet, linking bug reports to fix commits is important as it could shed light into the nature of bug fixing processes and expose patterns in software management. Bug linking solutions, such as ReLink, have been proposed. The demonstration of their effectiveness however faces a number of issues, including a reliability issue with their ground truth datasets as well as the extent of their measurements. We propose in this study a benchmark for evaluating bug linking solutions. This benchmark includes a dataset of about 12,000 bug links from 10 programs. These true links between bug reports and their fixes have been provided during bug fixing processes. We designed a number of research questions, to assess both quantitatively and qualitatively the effectiveness of a bug linking tool. Finally, we apply this benchmark on ReLink to report the strengths and limitations of this bug linking tool.

/pdf/empirical-evaluation-of-bug-linking-1uf2laajnc.pdf

Empirical Evaluation of Bug Linking

Pervasive computing constitutes a growing trend, aiming to embed smart devices into everyday objects. The limited resources of these devices and the ever-present need for lower production costs, lead to the research and development of lightweight cryptographic mechanisms. Block ciphers, the main symmetric key cryptosystems, perform well in this field. Nevertheless, stream ciphers are also relevant in ubiquitous computing applications, as they can be used to secure the communication in applications where the plaintext length is either unknown or continuous, like network streams. This paper provides the latest survey of stream ciphers for embedded systems. Lightweight implementations of stream ciphers in embedded hardware and software are examined as well as relevant authenticated encryption schemes. Their speed and simplicity enable compact and low-power implementations, allow them to excel in applications pertaining to resource-constrained devices. The outcomes of the International Organization for Standardization/International Electrotechnical Commission 29192-3 standard and the cryptographic competitions eSTREAM and Competition for Authenticated Encryption: Security, Applicability, and Robustness are summarized along with the latest results in the field. However, cryptanalysis has proven many of these schemes are actually insecure. From the 31 designs that are examined, only six of them have been found to be secure by independent cryptanalysis. A constrained benchmark analysis is performed on low-cost embedded hardware and software platforms. The most appropriate and secure solutions are then mapped in different types of applications. Copyright © 2015 John Wiley & Sons, Ltd.

A survey of lightweight stream ciphers for embedded systems

Nowadays social Networking and micro-blogging sites like Twitter are very popular and millions of users are registered on these websites. The users present on these website use these websites as a platform to express their thoughts and opinions. Our analysis of content posted on Twitter shows that users often post crime related information on Twitter. Among these crime related tweets some tweets are the crime messages that need police attention. Detection of such tweets can be beneficial in utilizing pattroling resources. The analysis of the data present on these websites can have an enormous impact. In this paper,the work is done on analyzing Twitter data to identify crime tweet that need police attention. Text mining based approach is used for classification of 369 tweets into crime and not-crime class. Classifiers such as Naive Bayesian, Random Forest, J48 and ZeroR are used. Among all of these four classifiers, Random forest classifier give the best accuracy of 98.1%.

/pdf/analysis-and-classification-of-crime-tweets-2jsmmvp1cl.pdf

Analysis and Classification of Crime Tweets

Defect tracking systems such as Bugzilla and JIRA and source code version control systems such as CVS and SVN are widely used applications to support software development and maintenance activities. Previous studies show that bug databases and version databases are often used as standalone and separate repositories without explicit linkages between issue reports and corresponding commit transactions. This is because developers often do not explicitly mention or tag commit transactions with the relevant bug report IDs. The lack of explicit links between these two databases has been identified as a serious process data quality issue (incomplete and biased data) having implications in predictive model building (such as defect density and error proneness computation) and hypothesis-testing based on the dataset. Researchers have proposed solutions to link the two databases and performed experiments on open source projects such as Fire Fox Mozilla. We review previous approaches and propose a novel technique (based on Fellegi-Sunter (FS) Model for record linkages) to automatically integrate the two databases that overcomes some of the drawbacks of traditional methods. We validate the proposed approach by performing experiments on publicly available bug and version dataset obtained from two open-source projects (Apache HTTP Server and WikiMedia). The results of our experiments demonstrate that the proposed solution is effective in recovering trace ability links (missing links) between bug-fixing commits and corresponding bug reports.

Applying Fellegi-Sunter (FS) Model for Traceability Link Recovery between Bug Databases and Version Archives

Asthma is a lung disease caused by the inflammation and narrowing of the airways that causes recurrent attacks of breathlessness and wheezing, and often can be life-threatening. Around 15–20 million people are suffering from asthma in India[1]. This paper aims at analyzing various data mining techniques for the prediction of asthma. The observations show that the fusion approach of naive bayes and neural network proved to be the best among classification algorithms in the diagnosis of asthma. This methodology is evaluated using 1024 raw data obtained from a city hospital. The proposed approach helps patients in their diagnosis of asthma.

Effective asthma disease prediction using naive Bayes — Neural network fusion technique

Bug reports submitted to an issue tracking system can belong to different categories such as crash, regression, security, cleanup, polish, performance and usability. A deeper understanding of the properties and features of various categories of bug reports can have implications in improving software maintenance processes, tools and practices. We identify several metrics and characteristics serving as dimensions on which various types of bug reports can be compared. We perform a case-study on Google Chromium Browser open-source project and conduct a series of experiments to calculate various metrics. We present a characterization study comparing different types of bug reports on metrics such as: statistics on close-time, number of stars, number of comments, discriminatory and frequent words for each class, entropy across reporters, entropy across component, opening and closing trend, continuity and debugging efficiency performance characteristics. The calculated metrics shows the similarities and differences on various dimensions for seven different types of bug reports.

Comparison of Seven Bug Report Types: A Case-Study of Google Chrome Browser Project

Bug or Fault localization is a process of identifying the specific location(s) or region(s) of source code (at various granularity levels such as the directory path, file, method or statement) that is faulty and needs to be modified to repair the defect. Fault localization is a routine task in software maintenance (corrective maintenance). Due to the increasing size and complexity of current software applications, automated solutions for fault localization can significantly reduce human effort and software maintenance cost.We present a technique (which falls into the class of static techniques for bug localization) for fault localization based on a character n-gram based Information Retrieval (IR) model. We frame the problem of bug localization as a relevant document(s) search task for a given query and investigate the application of character-level n-gram based textual features derived from bug reports and source-code file attributes. We implement the proposed IR model and evaluate its performance on dataset downloaded from two popular open-source projects (JBoss and Apache).We conduct a series of experiments to validate our hypothesis and present evidences to demonstrate that the proposed approach is effective. The accuracy of the proposed approach is measured in terms of the standard and commonly used SCORE and MAP (Mean Average Precision) metrics for the task of fault localization. Experimental results reveal that the median value for the SCORE metric for JBoss and Apache dataset is 99.03% and 93.70% respectively. We observe that for 16.16% of the bug reports in the JBoss dataset and for 10.67% of the bug reports in the Apache dataset, the average precision value (computed at all recall levels) is between 0.9 and 1.0.

Sangeeta Lal

Papers

Analysis and Classification of Crime Tweets

Applying Fellegi-Sunter (FS) Model for Traceability Link Recovery between Bug Databases and Version Archives

Effective asthma disease prediction using naive Bayes — Neural network fusion technique

Comparison of Seven Bug Report Types: A Case-Study of Google Chrome Browser Project

A static technique for fault localization using character n-gram based information retrieval model